There is a need to compensate for tool wear by continuous monitoring during precision machining. In general, in-process tool wear monitoring is carried out using the cutting force as the input signal. However, since the cutting condition varies for different machining operations, it is difficult to monitor the tool wear for general cases. Thus, to minimize the effect of the change of cutting force due to varying cutting conditions on tool wear monitoring, a force ratio is introduced as the input variable. The key point in this paper is to use both statistical and sensitivity analyses of the force ratio for the purpose of reliable tool wear monitoring. Also, to verify the results of the analysis, a neural network scheme is applied to perform one-step-ahead prediction of the flank wear from the force ratio obtained from a tool dynamometer. The results of statistical and sensitivity analyses based on the neural network approach correspond quite well to the results obtained from experiments.
All Science Journal Classification (ASJC) codes
- Ceramics and Composites
- Computer Science Applications
- Metals and Alloys
- Industrial and Manufacturing Engineering